creative-library-selector
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional, providing a structured algorithm and reference matrix for choosing front-end libraries (such as GSAP, Three.js, and Lenis). It contains no scripts, dynamic execution patterns, or obfuscated content.
- [INDIRECT_PROMPT_INJECTION]: The skill operates on user-provided project briefs and queries to generate technology recommendations. While this represents a data ingestion surface, the skill provides a rigid decision logic that mitigates the risk of arbitrary instruction following from processed content.
- Ingestion points: User queries regarding library choices and project briefs provided in
SKILL.md. - Boundary markers: None explicitly defined in the logic, but the structural nature of the 'Selection Algorithm' acts as a functional constraint.
- Capability inventory: The skill instructs the agent to generate dependency installation plans (e.g.,
npm install). The skill itself lacks file-write or network capabilities. - Sanitization: Standard LLM safety filters apply; the skill relies on the agent's internal reasoning to validate recommended package names.
Audit Metadata